Evgeny Kotkov <[email protected]> writes:
> Since this discussion has been lasting for over a month now, I proposed
> what I believe is the proper solution for the problem (and that's (1) )
> in /branches/1.9.x/STATUS.
After giving it more thought, I agree that changing other cases isn't required
and justified in context of this nomination. I undid the part of the change
touching other calling sites, except for those in dump.c that are required to
solve issue #4598, and updated the nomination.
This leaves us with a couple of related issues in Subversion 1.9.2 (fixed in
trunk), but we can handle them with separate nominations:
(1) svn_fs_contents_different() can produce invalid results under circumstances
described in issue 4554 [1]. In practice this means that, for instance,
one can have a working copy that misses an update in a file that happened
in the repository, and produces "Base checksum mismatch" errors on attempts
to commit changes in this file.
This issue was addressed by r1673875 [2] and follow-ups.
(2) svn_fs_props_different() relies on MD5 checksum for the comparison.
This issue was addressed by r1705638 [3] and follow-ups.
I'll look into doing that.
[1] https://issues.apache.org/jira/browse/SVN-4554
[2] https://svn.apache.org/r1673875
[3] https://svn.apache.org/r1705638
Regards,
Evgeny Kotkov